Is there a package that will do a cross-validatation with regularization for beta regression in R? I'm looking for an equivalent of glmnet for the betareg package.
I'm not sure whether there is a package that does exactly lasso estimation with CV like
glmnet does. However, for regularized regression with variable selection boosting with stability selection also often works well. This is implemented for beta regression in both
bamlss. The latter also contains a lasso regression term
la(). This is introduced in Groll at al. (2018). "Lasso-Type Penalization in the Framework of Generalized Additive Models for Location, Scale and Shape."RePEc:inn:wpaper:2018-16.